Craig's patch to the QuoteMeta function in pcrecpp.cc so that it escapes the

NUL character as backslash + 0 rather than backslash + NUL, because PCRE
doesn't support NULs in patterns.


git-svn-id: svn://vcs.exim.org/pcre/code/trunk@326 2f5784b3-3f2a-0410-8824-cb99058d5e15

4 files changed, 24 insertions, 10 deletions

diff --git a/ChangeLog b/ChangeLog
index 5cc4838..ddbe87b 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -29,6 +29,10 @@ Version 7.7 05-Mar-08
         
 5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
     --exclude_dir in the tests to avoid scanning .svn directories. 
+    
+6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
+    NUL character as backslash + 0 rather than backslash + NUL, because PCRE 
+    doesn't support NULs in patterns.
 
 
 Version 7.6 28-Jan-08
diff --git a/pcrecpp.cc b/pcrecpp.cc
index f34f0fc..52187a7 100644
--- a/pcrecpp.cc
+++ b/pcrecpp.cc
@@ -449,21 +449,27 @@ bool RE::Extract(const StringPiece& rewrite,
   // Note that it's legal to escape a character even if it has no
   // special meaning in a regular expression -- so this function does
   // that.  (This also makes it identical to the perl function of the
-  // same name; see `perldoc -f quotemeta`.)
+  // same name; see `perldoc -f quotemeta`.)  The one exception is
+  // escaping NUL: rather than doing backslash + NUL, like perl does,
+  // we do '\0', because pcre itself doesn't take embedded NUL chars.
   for (int ii = 0; ii < unquoted.size(); ++ii) {
     // Note that using 'isalnum' here raises the benchmark time from
     // 32ns to 58ns:
-    if ((unquoted[ii] < 'a' || unquoted[ii] > 'z') &&
-        (unquoted[ii] < 'A' || unquoted[ii] > 'Z') &&
-        (unquoted[ii] < '0' || unquoted[ii] > '9') &&
-        unquoted[ii] != '_' &&
-        // If this is the part of a UTF8 or Latin1 character, we need
-        // to copy this byte without escaping.  Experimentally this is
-        // what works correctly with the regexp library.
-        !(unquoted[ii] & 128)) {
+    if (unquoted[ii] == '\0') {
+      result += "\\0";
+    } else if ((unquoted[ii] < 'a' || unquoted[ii] > 'z') &&
+               (unquoted[ii] < 'A' || unquoted[ii] > 'Z') &&
+               (unquoted[ii] < '0' || unquoted[ii] > '9') &&
+               unquoted[ii] != '_' &&
+               // If this is the part of a UTF8 or Latin1 character, we need
+               // to copy this byte without escaping.  Experimentally this is
+               // what works correctly with the regexp library.
+               !(unquoted[ii] & 128)) {
       result += '\\';
+      result += unquoted[ii];
+    } else {
+      result += unquoted[ii];
     }
-    result += unquoted[ii];
   }
 
   return result;
diff --git a/pcrecpp.h b/pcrecpp.h
index a4638e1..a87a816 100644
--- a/pcrecpp.h
+++ b/pcrecpp.h
@@ -620,6 +620,9 @@ class PCRECPP_EXP_DEFN RE {
   //           1.5-2.0?
   // may become:
   //           1\.5\-2\.0\?
+  // Note QuoteMeta behaves the same as perl's QuoteMeta function,
+  // *except* that it escapes the NUL character (\0) as backslash + 0,
+  // rather than backslash + NUL.
   static string QuoteMeta(const StringPiece& unquoted);
 
 
diff --git a/pcrecpp_unittest.cc b/pcrecpp_unittest.cc
index 2e5f2df..44e0cc9 100644
--- a/pcrecpp_unittest.cc
+++ b/pcrecpp_unittest.cc
@@ -497,6 +497,7 @@ static void TestQuotaMetaSimple() {
   TestQuoteMeta("((a|b)c?d*e+[f-h]i)");
   TestQuoteMeta("((?!)xxx).*yyy");
   TestQuoteMeta("([");
+  TestQuoteMeta(string("foo\0bar", 7));
 }
 
 static void TestQuoteMetaSimpleNegative() {